How to select the detected objects class with torch.hub?

[tgrmev]

I am calling yolo yolov5m.pt model with this code. I only want to detect the people from my environment. For this aim what should I add to this line ?
model = torch.hub.load("./yolov5", 'custom', path="./yolov5/yolov5m.pt", source='local')

I tried with this : model = torch.hub.load("./yolov5", 'custom', path="./yolov5/yolov5m.pt", source='local' , classes = 0 ) and then I got TypeError: custom() got an unexpected keyword argument 'classes' error. Is it not possible to select the detected class, if we work with custom model?

[Robotatron]

Do a prediction first, then select your class.

[tgrmev]

Could you write the line of that code after model = torch.hub.load("./yolov5", 'custom', path="./yolov5/yolov5m.pt", source='local') ?